# Generalized Linear Models in Spark MLlib and SparkR by Xiangrui Meng

date post

21-Apr-2017Category

## Data & Analytics

view

1.776download

0

Embed Size (px)

### Transcript of Generalized Linear Models in Spark MLlib and SparkR by Xiangrui Meng

Generalized Linear Models in Spark MLlib and SparkR

Xiangrui Meng joint with Joseph Bradley, Eric Liang, Yanbo Liang (MiningLamp), DB Tsai (Netflix), et al.

2016/02/17 - Spark Summit East

About me

Software Engineer at Databricks Spark PMC member and MLlib/PySpark maintainer Ph.D. from Stanford on randomized algorithms for large-

scale linear regression problems

2

Outline

Generalized linear models (GLMs) linear regression / logistic regression / general form accelerated failure time (AFT) model for survival analysis intercept / regularization / weights

GLMs in MLlib and SparkR demo: R formula in Spark

Implementing GLMs gradient descent / L-BFGS / OWL-QN weighted least squares / iteratively re-weighted least squares (IRLS) performance tips

3

Generalized linear models

Linear regression

5

https://en.wikipedia.org/wiki/Linear_regression

inference / prediction

https://en.wikipedia.org/wiki/Linear_regression

Linear least squares

m observations: x: explanatory variables, y: dependent variable assumes linear relationship between x and y

minimizes the sum of the squares of the errors

6

Linear least squares

the oldest linear model, trace back to Gauss the simplest and the most studied linear model has analytic solutions easy to solve easy to inspect

sensitive to outliers

7

Logistic regression

8

https://en.wikipedia.org/wiki/Logistic_regression

https://en.wikipedia.org/wiki/Logistic_regression

Logistic regression

classification with binary response: true/false, clicked/not clicked, liked/disliked

uses logistic function to indicate the likelihood

maximizes the sum of the log-likelihoods, i.e.,

9

Logistic regression

one of the simplest binary classification models widely used in industry relatively easy to solve easy to interpret

10

Multinomial logistic regression

classification with multiclass response: uses softmax function to indicate likelihood

maximizes the sum of log-likelihoods

11

Generalized linear models (GLMs)

Both linear least squares and logistic regression are special cases of generalized linear models. A GLM is specified by the following: a distribution of the response (from the exponential family), a link function g such that

maximizes the sum of log-likelihoods

12

Distributions and link functions

13

Model Distribution Link

linear least squares normal identity

logistic regression binomial logit

multinomial logic regression multinomial generalized logit

Poisson regression Poisson log

gamma regression gamma inverse

Accelerated failure time (AFT) model

m observations: y: survival time, c: censor variable (alive or dead) assumes the effect of an explanatory variable is to

accelerate or decelerate the life time by some constant uses maximum likelihood estimation while treating

censored and uncensored observations differently

14

AFT model for survival analysis

one popular parametric model for survival analysis widely used for lifetime estimation and churn analysis could be solved under the same framework as GLMs

15

Intercept, regularization, and weights

In practice, a linear model is often more complexwhere w describes instance weights, beta_0 is the intercept term to adjust bias, and sigma regularized beta with a constant lambda > 0 to avoid overfitting.

16

Types of regularization

Ridge (L2): easy to solve (strongly convex)

Lasso (L1): enhance model sparsity harder to solver (though still convex)

Elastic-Net: Others: group lasso, nonconvex, etc

17

GLMs in MLlib and SparkR

GLMs in Spark MLlib

Linear models in MLlib are implemented as ML pipeline estimators. They accept the following params: featuresCol: a vector column containing features (x) labelCol: a double column containing responses (y) weightCol: a double column containing weights (w) regType: regularization type, none, l1, l2, elastic-net regParam: regularization constant fitIntercept: whether to fit an intercept term

19

Fit a linear model in MLlib

from pyspark.ml.classification import LogisticRegression

# Load training data training = sqlContext.read.parquet(path/to/training)

lr = LogisticRegression( weightCol=weight, fitIntercept=False, maxIter=10, regParam=0.3, elasticNetParam=0.8)

# Fit the model model = lr.fit(training)

20

Make predictions and evaluate models

from pyspark.ml.evaluation import BinaryClassificationEvaluator

test = sqlContext.read.parquet(path/to/test)

# make predictions by calling transform predictions = model.tranform(test)

# create a binary classification evaluator evaluator = BinaryClassificationEvaluator( metricName=areaUnderROC) evaulator.evaluate(predictions)

21

GLMs in SparkR

In Python/Scala/Java, we keep the APIs about the same for consistency. But in SparkR, we make the APIs similar to existing ones in R (or R packages).

# Create the DataFrame df

R formula in SparkR

23

R provides model formula to express linear models. We support the following R formula operators in SparkR:

`~` separate target and terms

`+` concat terms, "+ 0" means removing intercept

`-` remove a term, "- 1" means removing intercept `:` interaction (multiplication for numeric values, or binarized categorical

values)

`.` all columns except target

For example, y ~ x + z + x:z -1 means using x, z, and their interaction (x:z) to predict y without intercept (-1).

Demo: GLMs in Spark

using Databricks Community Edition!

https://databricks-prod-cloudfront.cloud.databricks.com/public/4027ec902e239c93eaaa8714f173bcfc/1233855/468463116641886/588180/80fb1ef84e.html

Implementing GLMs

Row-based distributed storage

26

w x y

w x y

partition 1

partition 2

Gradient descent methods

Stochastic gradient descent (SGD): trade-offs on the merge scheme and convergence

Mini-batch SGD: hard to sample mini-batches efficiently communication overhead on merging gradients

Batch gradient descent: slow convergence

27

Quasi-Newton methods

Newtons method converges much than GD, but it requires second-order information: L-BFGS works for smooth objectives. It approximates the

inverse Hessian using only first-order information. OWL-QN works for objectives with L1 regularization. MLlib calls L-BFGS/OWL-QN implemented in breeze.

28

Direct methods for linear least squares

Linear least squares has an analytic solution:

The solution could be computed directly or through QR factorization, both of which are implemented in Spark. requires only a single pass efficient when the number of features is small (

Iteratively re-weighted least squares (IRLS)

Generalized linear models with exponential family can be solved via iteratively re-weighted least squares (IRLS). linearizes the objective at the current solution solves the weighted linear least squares problem repeat above steps until convergence

efficient when the number of features is small (

Verification using R

Besides normal tests, we also verify our implementation using R.

/* df

Standardization

To match the result in both R and glmnet, the most popular R package for GLMs, we provide options to standardize features and labels before training:where delta is the stddev of labels, and sigma_j is the stddev of the j-th feature column.

32

Performance tips

Utilize sparsity. Use tree aggregation and torrent broadcast. Watch numerical issues, e.g., log(1+exp(x)). Do not change input data. Scaling could be applied after

each iteration and intercept could be derived later.

33

Future directions

easy handling of categorical features and labels better R formula support more model summary statistics feature parity in different languages model parallelism vector-free L-BFGS with 2D partitioning (WIP)

using matrix kernels

34

Other GLM implementations on Spark

CoCoA+: communication-efficient optimization LIBLINEAR for Spark: a Spark port of LIBLINEAR sparkGLM: an R-like GLM package for Spark TFOCS for Spark: first-order conic solvers for Spark General-purpose packages that implement GLMs aerosolve, DistML, sparkling-water, thunder, zen, etc

and more on Spark Packages

35

https://github.com/gingsmith/cocoahttps://www.csie.ntu.edu.tw/~cjlin/libsvmtools/distributed-liblinear/https://github.com/AlteryxLabs/sparkGLMhttps://github.com/databricks/spark-tfocshttps://github.com/airbnb/aerosolvehttps://github.com/intel-machine-learning/DistMLhttps://github.com/h2oai/sparkling-waterhttps://github.com/thunder-project/thunderhttps://github.com/cloudml/zenhttp://spark-packages.org

Thank you. MLlib user guide and roadmap for Spark 2.0 GLMs on Wikipedia Databricks Community Edition, blog posts, and careers

http://spark.apache.org/docs/latest/ml-guide.htmlhttps://issues.apache.org/jira/browse/SPARK-12626https://en.wikipedia.org/wiki/Generalized_linear_modelhttps://community.cloud.databricks.com/https://databricks.com/bloghttps://databricks.com/company/career